Evaluating Messaging on Prenatal Health Behaviors Using Social Media Data: Systematic Review

BACKGROUND

Social media platforms are increasingly being used to disseminate messages about prenatal health. However, to date, we lack a systematic assessment of how to evaluate the impact of official prenatal health messaging and campaigns using social media data.

OBJECTIVE

This study aims to review both the published and gray literature on how official prenatal health messaging and campaigns have been evaluated to date in terms of impact, acceptability, effectiveness, and unintended consequences, using social media data.

METHODS

A total of 6 electronic databases were searched and supplemented with the hand-searching of reference lists. Both published and gray literature were eligible for review. Data were analyzed using content analysis for descriptive data and a thematic synthesis approach to summarize qualitative evidence. A quality appraisal tool, designed especially for use with social media data, was used to assess the quality of the included articles.

RESULTS

A total of 11 studies were eligible for the review. The results showed that the most common prenatal health behavior targeted was alcohol consumption, and Facebook was the most commonly used source of social media data. The majority (n=6) of articles used social media data for descriptive purposes only. The results also showed that there was a lack of evaluation of the effectiveness, acceptability, and unintended consequences of the prenatal health message or campaign.

CONCLUSIONS

Social media is a widely used and potentially valuable resource for communicating and evaluating prenatal health messaging. However, this review suggests that there is a need to develop and adopt sound methodology on how to evaluate prenatal health messaging using social media data, for the benefit of future research and to inform public health practice.

Identifying mental health discussion topic in social media community: subreddit of bipolar disorder analysis

Online platforms allow individuals to connect with others, share experiences, and find communities with similar interests, providing a sense of belonging and reducing feelings of isolation. Numerous previous studies examined the content of online health communities to gain insights into the sentiments surrounding mental health conditions. However, there is a noticeable gap in the research landscape, as no study has specifically concentrated on conducting an in-depth analysis or providing a comprehensive visualization of Bipolar disorder. Therefore, this study aimed to address this gap by examining the Bipolar subreddit online community, where we collected 1,460,447 posts as plain text documents for analysis. By employing LDA topic modeling and sentiment analysis, we found that the Bipolar disorder online community on Reddit discussed various aspects of the condition, including symptoms, mood swings, diagnosis, and medication. Users shared personal experiences, challenges, and coping strategies, seeking support and connection. Discussions related to therapy and medication were prevalent, emphasizing the importance of finding suitable therapists and managing medication side effects. The online community serves as a platform for seeking help, advice, and information, highlighting the role of social support in managing bipolar disorder. This study enhances our understanding of individuals living with bipolar disorder and provides valuable insights and feedback for researchers developing mental health interventions.

A Novel Foodborne Illness Detection and Web Application Tool Based on Social Media

Foodborne diseases and outbreaks are significant threats to public health, resulting in millions of illnesses and deaths worldwide each year. Traditional foodborne disease surveillance systems rely on data from healthcare facilities, laboratories, and government agencies to monitor and control outbreaks. Recently, there is a growing recognition of the potential value of incorporating social media data into surveillance systems. This paper explores the use of social media data as an alternative surveillance tool for foodborne diseases by collecting large-scale Twitter data, building food safety data storage models, and developing a novel frontend foodborne illness surveillance system. Descriptive and predictive analyses of the collected data were conducted in comparison with ground truth data reported by the U.S. Centers for Disease Control and Prevention (CDC). The results indicate that the most implicated food categories and the distributions from both Twitter and the CDC were similar. The system developed with Twitter data could complement traditional foodborne disease surveillance systems by providing near-real-time information on foodborne illnesses, implicated foods, symptoms, locations, and other information critical for detecting a potential foodborne outbreak.

Development and Assessment of a Social Media-Based Construct of Firearm Ownership: Computational Derivation and Benchmark Comparison

BACKGROUND

Gun violence research is characterized by a dearth of data available for measuring key constructs. Social media data may offer a potential opportunity to significantly reduce that gap, but developing methods for deriving firearms-related constructs from social media data and understanding the measurement properties of such constructs are critical precursors to their broader use.

OBJECTIVE

This study aimed to develop a machine learning model of individual-level firearm ownership from social media data and assess the criterion validity of a state-level construct of ownership.

METHODS

We used survey responses to questions on firearm ownership linked with Twitter data to construct different machine learning models of firearm ownership. We externally validated these models using a set of firearm-related tweets hand-curated from the Twitter Streaming application programming interface and created state-level ownership estimates using a sample of users collected from the Twitter Decahose application programming interface. We assessed the criterion validity of state-level estimates by comparing their geographic variance to benchmark measures from the RAND State-Level Firearm Ownership Database.

RESULTS

We found that the logistic regression classifier for gun ownership performs the best with an accuracy of 0.7 and an F₁-score of 0.69. We also found a strong positive correlation between Twitter-based estimates of gun ownership and benchmark ownership estimates. For states meeting a threshold requirement of a minimum of 100 labeled Twitter users, the Pearson and Spearman correlation coefficients are 0.63 (P<.001) and 0.64 (P<.001), respectively.

CONCLUSIONS

Our success in developing a machine learning model of firearm ownership at the individual level with limited training data as well as a state-level construct that achieves a high level of criterion validity underscores the potential of social media data for advancing gun violence research. The ownership construct is an important precursor for understanding the representativeness of and variability in outcomes that have been the focus of social media analyses in gun violence research to date, such as attitudes, opinions, policy stances, sentiments, and perspectives on gun violence and gun policy. The high criterion validity we achieved for state-level gun ownership suggests that social media data may be a useful complement to traditional sources of information on gun ownership such as survey and administrative data, especially for identifying early signals of changes in geographic patterns of gun ownership, given the immediacy of the availability of social media data, their continuous generation, and their responsiveness. These results also lend support to the possibility that other computationally derived, social media-based constructs may be derivable, which could lend additional insight into firearm behaviors that are currently not well understood. More work is needed to develop other firearms-related constructs and to assess their measurement properties.

Increasing biodiversity knowledge through social media: A case study from tropical Bangladesh

Citizen science programs are becoming increasingly popular among naturalists but remain heavily biased taxonomically and geographically. However, with the explosive popularity of social media and the near-ubiquitous availability of smartphones, many post wildlife photographs on social media. Here, we illustrate the potential of harvesting these data to enhance our biodiversity understanding using Bangladesh, a tropical biodiverse country, as a case study. We compared biodiversity records extracted from Facebook with those from the Global Biodiversity Information Facility (GBIF), collating geospatial records for 1013 unique species, including 970 species from Facebook and 712 species from GBIF. Although most observation records were biased toward major cities, the Facebook records were more evenly spatially distributed. About 86% of the Threatened species records were from Facebook, whereas the GBIF records were almost entirely Of Least Concern species. To reduce the global biodiversity data shortfall, a key research priority now is the development of mechanisms for extracting and interpreting social media biodiversity data.

Social media data shed light on air-conditioning interest of heat-vulnerable regions and sociodemographic groups

Cooling homes with air conditioners is a vital adaptation approach, but the wider adoption of air conditioners can increase hydrofluorocarbon emissions that have high global warming potential and carbon emissions as a result of more fossil energy consumption. The scale and scope of future cooling demand worldwide are, however, uncertain because the extent and drivers of air-conditioning adoption remain unclear. Here, using 2021 and 2022 Facebook and Instagram data from 113 countries, we investigate the usability of social media advertising data to address these data gaps in relation to the drivers of air-conditioning adoption. We find that social media data might represent air-conditioning purchasing trends. Globally, parents of small children and middle-aged, highly educated married or cohabiting males tend to express greater interest in air-conditioning adoption. In regions with high heat vulnerability yet little empirical data on cooling demand (e.g., the Middle East and North Africa), these sociodemographic factors play a more prominent role. These findings can strengthen our understanding of future cooling demand for more sustainable cooling management.

Public Opinions on Stray Cats in China, Evidence from Social Media Data

The management of stray cats is often contentious because public perceptions about these animals are different. Using user-generated content from Weibo, this study investigated Chinese citizens' opinions on stray cats on a large scale. Through the techniques of natural language processing, we obtained each Weibo post's topics and sentiment propensity. The results showed that: (1) there were some irresponsible feeding behaviors among citizens; (2) public perceptions of the ecological impacts caused by stray cats were unlike; (3) the trap-neuter-return (TNR) method served high support in public discussion; (4) knowledge about stray cats' ecological impacts was positively correlated with support for the lethal control methods in management. Based on these findings, we suggested that management policies should be dedicated to (1) communicating to the (potential) cat feeders about the negative aspects of irresponsible feeding behaviors; (2) raising "ecological awareness" campaigns for the public as well as highlighting the environmental impacts caused by stray cats; (3) understanding citizens' perceptions toward different management scenarios and making decisions accordingly. In addition, this study also suggested that social media data can provide useful information about people's opinions on wild animals and their management. Policies would benefit by taking this source of information into the decision-making process.

Estimating County-Level Overdose Rates Using Opioid-Related Twitter Data: Interdisciplinary Infodemiology Study

BACKGROUND

There were an estimated 100,306 drug overdose deaths between April 2020 and April 2021, a three-quarter increase from the prior 12-month period. There is an approximate 6-month reporting lag for provisional counts of drug overdose deaths from the National Vital Statistics System, and the highest level of geospatial resolution is at the state level. By contrast, public social media data are available close to real-time and are often accessible with precise coordinates.

OBJECTIVE

The purpose of this study is to assess whether county-level overdose mortality burden could be estimated using opioid-related Twitter data.

METHODS

International Classification of Diseases (ICD) codes for poisoning or exposure to overdose at the county level were obtained from CDC WONDER. Demographics were collected from the American Community Survey. The Twitter Application Programming Interface was used to obtain tweets that contained any of the 36 terms with drug names. An unsupervised classification approach was used for clustering tweets. Population-normalized variables and polynomial population-normalized variables were produced. Furthermore, z scores of the Getis Ord Gi clustering statistic were produced, and both these scores and their polynomial counterparts were explored in regression modeling of county-level overdose mortality burden. A series of linear regression models were used for predictive modeling to explore the interpretability of the analytical output.

RESULTS

Modeling overdose mortality with normalized demographic variables alone explained only 7.4% of the variability in county-level overdose mortality, whereas this was approximately doubled by the use of specific demographic and Twitter data covariates based on a backward selection approach. The highest adjusted R² and lowest AIC (Akaike Info Criterion) were obtained for the model with normalized demographic variables, normalized z scores from geospatial analyses, and normalized topic counts (adjusted R²=0.133, AIC=8546.8). The z scores of the Getis Ord Gi statistic appeared to have improved utility over population-normalization alone. In this model, median age, female population, and tweets about web-based drug sales were positively associated with opioid mortality. Asian race and Hispanic ethnicity were significantly negatively associated with county-level burdens of overdose mortality.

CONCLUSIONS

Social media data, when transformed using certain statistical approaches, may add utility to the goal of producing closer to real-time county-level estimates of overdose mortality. Prediction of opioid-related outcomes can be advanced to inform prevention and treatment decisions. This interdisciplinary approach can facilitate evidence-based funding decisions for various substance use disorder prevention and treatment programs.

Sharing social media data: The role of past experiences, attitudes, norms, and perceived behavioral control

Social media data (SMD) have become an important data source in the social sciences. The purpose of this paper is to investigate the experiences and practices of researchers working with SMD in their research and gain insights into researchers' sharing behavior and influencing factors for their decisions. To achieve these aims, we conducted a survey study among researchers working with SMD. The questionnaire covered different topics related to accessing, (re)using, and sharing SMD. To examine attitudes toward data sharing, perceived subjective norms, and perceived behavioral control, we used questions based on the Theory of Planned Behavior (TPB). We employed a combination of qualitative and quantitative analyses. The results of the qualitative analysis show that the main reasons for not sharing SMD were that sharing was not considered or needed, as well as legal and ethical challenges. The quantitative analyses reveal that there are differences in the relative importance of past sharing and reuse experiences, experienced challenges, attitudes, subjective norms, and perceived behavioral control as predictors of future SMD sharing intentions, depending on the way the data should be shared (publicly, with restricted access, or upon personal request). Importantly, the TPB variables have predictive power for all types of SMD sharing.

The Evaluation of Rural Outdoor Dining Environment from Consumer Perspective

The quality of the environment should be measured by the satisfaction of the public and guided by the issues of public concern. With the development of the internet, social media as the main platform for people to exchange information has become a data source for planning and management analysis. Nowadays, the rural catering industry is becoming increasingly competitive, especially after the pandemic. How to further enhance the competitiveness of the rural catering industry has become a hot topic in the industry. From the perspective of consumers, we explored consumers' preferences in a rural outdoor dining environment through social media data. The research analyzed the social media data through manual collection and object detection, divided the landscape of the rural outdoor dining environment into eight categories with 35 landscape elements, and then used BP (Back Propagation) neural network nonlinear fitting and least square linear fitting to analyze the 11,410 effective review pictures from eight rural restaurants' social media comments in Chengdu. We derived the degree of consumer preference for the landscape quality of the rural outdoor dining environment and analyzed the differences in preference among three different groups (regular customers, customers with children, and customers with the elderly). The study found that agricultural resources are an important factor in the competitiveness of rural restaurant environments; that children's emotions when using activity facilities can positively influence consumers' dining experiences; that safety and hygiene environment are important factors influencing the decisions of parent-child dining; and that older people are more interested in outdoor nature, etc. The research results provide suggestions and knowledge for rural restaurant managers and designers through human-oriented needs from the perspective of consumers, and clarify the preferences and expectations of different consumer groups for rural restaurant landscapes while achieving the goal of rural landscape protection.

Spatial-Temporal Pattern Evolution of Public Sentiment Responses to the COVID-19 Pandemic in Small Cities of China: A Case Study Based on Social Media Data Analysis

The impact of the COVID-19 pandemic on public mental health has become increasingly prominent. Therefore, it is of great value to study the spatial-temporal characteristics of public sentiment responses to COVID-19 exposure to improve urban anti-pandemic decision-making and public health resilience. However, the majority of recent studies have focused on the macro scale or large cities, and there is a relative lack of adequate research on the small-city scale in China. To address this lack of research, we conducted a case study of Shaoxing city, proposed a spatial-based pandemic-cognition-sentiment (PCS) conceptual model, and collected microblog check-in data and information on the spatial-temporal trajectory of cases before and after a wave of the COVID-19 pandemic. The natural language algorithm of dictionary-based sentiment analysis (DSA) was used to calculate public sentiment strength. Additionally, local Moran's I, kernel-density analysis, Getis-Ord Gi* and standard deviation ellipse methods were applied to analyze the nonlinear evolution and clustering characteristics of public sentiment spatial-temporal patterns at the small-city scale concerning the pandemic. The results reveal that (1) the characteristics of pandemic spread show contagion diffusion at the micro level and hierarchical diffusion at the macro level, (2) the pandemic has a depressive effect on public sentiment in the center of the outbreak, and (3) the pandemic has a nonlinear gradient negative impact on mood in the surrounding areas. These findings could help propose targeted pandemic prevention policies applying spatial intervention to improve residents' mental health resilience in response to future pandemics.

Extraction of Explicit and Implicit Cause-Effect Relationships in Patient-Reported Diabetes-Related Tweets From 2017 to 2021: Deep Learning Approach

Background

Intervening in and preventing diabetes distress requires an understanding of its causes and, in particular, from a patient’s perspective. Social media data provide direct access to how patients see and understand their disease and consequently show the causes of diabetes distress.

Objective

Leveraging machine learning methods, we aim to extract both explicit and implicit cause-effect relationships in patient-reported diabetes-related tweets and provide a methodology to better understand the opinions, feelings, and observations shared within the diabetes online community from a causality perspective.

Methods

More than 30 million diabetes-related tweets in English were collected between April 2017 and January 2021. Deep learning and natural language processing methods were applied to focus on tweets with personal and emotional content. A cause-effect tweet data set was manually labeled and used to train (1) a fine-tuned BERTweet model to detect causal sentences containing a causal relation and (2) a conditional random field model with Bidirectional Encoder Representations from Transformers (BERT)-based features to extract possible cause-effect associations. Causes and effects were clustered in a semisupervised approach and visualized in an interactive cause-effect network.

Results

Causal sentences were detected with a recall of 68% in an imbalanced data set. A conditional random field model with BERT-based features outperformed a fine-tuned BERT model for cause-effect detection with a macro recall of 68%. This led to 96,676 sentences with cause-effect relationships. “Diabetes” was identified as the central cluster followed by “death” and “insulin.” Insulin pricing–related causes were frequently associated with death.

Conclusions

A novel methodology was developed to detect causal sentences and identify both explicit and implicit, single and multiword cause, and the corresponding effect, as expressed in diabetes-related tweets leveraging BERT-based architectures and visualized as cause-effect network. Extracting causal associations in real life, patient-reported outcomes in social media data provide a useful complementary source of information in diabetes research.

China's Mismatch of Public Awareness and Biodiversity Threats under Economic Trade

China is one of the countries with high biodiversity on the globe, but suffers extreme biodiversity loss due to the increasingly interconnected economy. Understanding the nation-level public awareness of biodiversity under economic trades is crucial for implementing sustainable production and consumption of Sustainable Development Goals (SDGs). This study is the first to assess the public awareness of biodiversity loss associated with China's interprovincial trades by utilizing social media data and the multiregion input-output (MRIO) table. Results show that China's interprovincial trades cause heavy threats not only to local species but to distant species. Approximately 60% of provinces displace over half of their consumption-based biodiversity threats to other provinces. Nevertheless, individuals do not clearly realize their responsibility for the distant biodiversity they consumed, with a large mismatch both in popularity (Gini index = 0.51, Robin index = 39.57) and donation (Gini index = 0.69, Robin index = 54.58). To alleviate this phenomenon, our analysis suggests that the expansion of national-level nature reserves may be effectively beneficial to public biodiversity awareness, showing significantly positive partial correlation coefficients with individuals' popularity and donations. These insights provided by this study offer targeted information for conservation and call for synergistic collaboration between the civil society, especially consumers, and governments to turn the tide of biodiversity loss.

Determining containment policy impacts on public sentiment during the pandemic using social media data

SignificanceFor effective pandemic response, policymakers need tools that can assess policy impacts in near real-time. This requires policymakers to monitor changes in public well-being due to policy interventions. Particularly, containment measures affect people's mental well-being, yet changes in public emotions and sentiments are challenging to assess. Our work provides a solution by using social media posts to compute salient concerns and daily public sentiment values as a proxy of mental well-being. We demonstrate how public sentiment and concerns are impacted by various containment policy sub-types. This approach provides key benefits of using a data-driven approach to identify public concerns and provides near real-time assessment of policy impacts by computing daily public sentiment based on postings on social media.

Correlation Analysis between Urban Elements and COVID-19 Transmission Using Social Media Data

The outbreak of the COVID-19 has become a worldwide public health challenge for contemporary cities during the background of globalization and planetary urbanization. However, spatial factors affecting the transmission of the disease in urban spaces remain unclear. Based on geotagged COVID-19 cases from social media data in the early stage of the pandemic, this study explored the correlation between different infectious outcomes of COVID-19 transmission and various factors of the urban environment in the main urban area of Wuhan, utilizing the multiple regression model. The result shows that most spatial factors were strongly correlated to case aggregation areas of COVID-19 in terms of population density, human mobility and environmental quality, which provides urban planners and administrators valuable insights for building healthy and safe cities in an uncertain future.

Modelling the geographical spread of HIV among MSM in Guangdong, China: a metapopulation model considering the impact of pre-exposure prophylaxis

Men who have sex with men (MSM) make up the majority of new human immunodeficiency virus (HIV) diagnoses among young people in China. Understanding HIV transmission dynamics among the MSM population is, therefore, crucial for the control and prevention of HIV infections, especially for some newly reported genotypes of HIV. This study presents a metapopulation model considering the impact of pre-exposure prophylaxis (PrEP) to investigate the geographical spread of a hypothetically new genotype of HIV among MSM in Guangdong, China. We use multiple data sources to construct this model to characterize the behavioural dynamics underlying the spread of HIV within and between 21 prefecture-level cities (i.e. Guangzhou, Shenzhen, Foshan, etc.) in Guangdong province: the online social network via a gay social networking app, the offline human mobility network via the Baidu mobility website, and self-reported sexual behaviours among MSM. Results show that PrEP initiation exponentially delays the occurrence of the virus for the rest of the cities transmitted from the initial outbreak city; hubs on the movement network, such as Guangzhou, Shenzhen, and Foshan are at a higher risk of 'earliest' exposure to the new HIV genotype; most cities acquire the virus directly from the initial outbreak city while others acquire the virus from cities that are not initial outbreak locations and have relatively high betweenness centralities, such as Guangzhou, Shenzhen and Shantou. This study provides insights in predicting the geographical spread of a new genotype of HIV among an MSM population from different regions and assessing the importance of prefecture-level cities in the control and prevention of HIV in Guangdong province. This article is part of the theme issue 'Data science approach to infectious disease surveillance'.

Mapping Total Exceedance PM2.5 Exposure Risk by Coupling Social Media Data and Population Modeling Data

The PM2.5 exposure risk assessment is the foundation to reduce its adverse effects. Population survey-related data have been deficient in high spatiotemporal detailed descriptions. Social media data can quantify the PM2.5 exposure risk at high spatiotemporal resolutions. However, due to the no-sample characteristics of social media data, PM2.5 exposure risk for older adults is absent. We proposed combining social media data and population survey-derived data to map the total PM2.5 exposure risk. Hourly exceedance PM2.5 exposure risk indicators based on population modeling (HEPEpmd) and social media data (HEPEsm) were developed. Daily accumulative HEPEsm and HEPEpsd ranged from 0 to 0.009 and 0 to 0.026, respectively. Three peaks of HEPEsm and HEPEpsd were observed at 13:00, 18:00, and 22:00. The peak value of HEPEsm increased with time, which exhibited a reverse trend to HEPEpsd. The spatial center of HEPEsm moved from the northwest of the study area to the center. The spatial center of HEPEpsd moved from the northwest of the study area to the southwest of the study area. The expansion area of HEPEsm was nearly 1.5 times larger than that of HEPEpsd. The expansion areas of HEPEpsd aggregated in the old downtown, in which the contribution of HEPEpsd was greater than 90%. Thus, this study introduced various source data to build an easier and reliable method to map total exceedance PM2.5 exposure risk. Consequently, exposure risk results provided foundations to develop PM2.5 pollution mitigation strategies as well as scientific supports for sustainability and eco-health achievement.

Exploring the Influencing Factors of the Recreational Utilization and Evaluation of Urban Ecological Protection Green Belts for Urban Renewal: A Case Study in Shanghai

With the continuous expansion of urban construction land, the green belts aiming for ecological protection have ensured a sustainable and effective function of regional ecosystem services. At the same time, these ecological green belts are expected to develop their compound service potentials with the development of cities. In order to meet the increasing demand of urban residents for the recreational utilization of urban green space, the primary function of the ecological green belts has transformed from being purely ecological to a combination of being ecological and recreational. Based on social media data, which has the characteristics of a large amount of accessible geographic information, this study used multiple regression models to analyze the recreational utilization intensity of ecological protection green belts with a case study in the green belt of Shanghai, China. The research results showed that the internal elements (total external area, water area, etc.) of the Shanghai green belt have positive correlations with its recreational utilization. The impact of external factors was inconclusive on the recreational utilization of the outer forest belt (the number of subway stations in accessibility factors was negatively correlated; the number of cultural facilities and the number of restaurants in the surrounding service facilities were positively related). Combined with the "Shanghai City Master Plan (2017-2035)", this study suggests potential zones for the recreational transformation of the Shanghai green belt, provides a theoretical and practical basis for improving the recreational utilization of an urban ecological protection green belt and contributes to the sustainable development of ecological protection green belts in high-density cities.

A biomedically oriented automatically annotated Twitter COVID-19 dataset

The use of social media data, like Twitter, for biomedical research has been gradually increasing over the years. With the coronavirus disease 2019 (COVID-19) pandemic, researchers have turned to more non-traditional sources of clinical data to characterize the disease in near-real time, study the societal implications of interventions, as well as the sequelae that recovered COVID-19 cases present. However, manually curated social media datasets are difficult to come by due to the expensive costs of manual annotation and the efforts needed to identify the correct texts. When datasets are available, they are usually very small and their annotations don't generalize well over time or to larger sets of documents. As part of the 2021 Biomedical Linked Annotation Hackathon, we release our dataset of over 120 million automatically annotated tweets for biomedical research purposes. Incorporating best-practices, we identify tweets with potentially high clinical relevance. We evaluated our work by comparing several SpaCy-based annotation frameworks against a manually annotated gold-standard dataset. Selecting the best method to use for automatic annotation, we then annotated 120 million tweets and released them publicly for future downstream usage within the biomedical domain.

A Biomedically oriented automatically annotated Twitter COVID-19 Dataset

The use of social media data, like Twitter, for biomedical research has been gradually increasing over the years. With the COVID-19 pandemic, researchers have turned to more nontraditional sources of clinical data to characterize the disease in near real-time, study the societal implications of interventions, as well as the sequelae that recovered COVID-19 cases present (Long-COVID). However, manually curated social media datasets are difficult to come by due to the expensive costs of manual annotation and the efforts needed to identify the correct texts. When datasets are available, they are usually very small and their annotations do not generalize well over time or to larger sets of documents. As part of the 2021 Biomedical Linked Annotation Hackathon, we release our dataset of over 120 million automatically annotated tweets for biomedical research purposes. Incorporating best practices, we identify tweets with potentially high clinical relevance. We evaluated our work by comparing several SpaCy-based annotation frameworks against a manually annotated gold-standard dataset. Selecting the best method to use for automatic annotation, we then annotated 120 million tweets and released them publicly for future downstream usage within the biomedical domain.

"Leaving Was a Process, Not an Event": The Lived Experience of Dating and Domestic Violence in 140 Characters

One of the most frequent refrains heard in the public discourse on intimate partner violence (IPV) is why do they stay? The literature has demonstrated that IPV victims face multiple barriers to safely exiting their relationships. Currently, there has been a limited examination of the role social media can play in elucidating the lived experience of IPV. With 25% of the population using Twitter, there are opportunities to examine its utility for deepening understandings of IPV. Using data generated from the #WhyIStayed Twitter campaign, the purpose of this study is to examine Twitter users' reasons for staying in their abusive relationships. The study sample (n = 3,086) is composed of a random sample of 61,725 English speaking tweets globally that employed the #WhyIStayed and #WhyILeft hashtags. We analyzed all tweets using thematic content analysis methods. This process involved multiple rounds of coding. In response to #WhyIStayed, Tweeters worldwide shared the barriers they faced that made leaving their abusive partners difficult. Seven primary themes emerged that influenced their decision-making processes: (a) impact of IPV on personal well-being, (b) lack of awareness regarding the dynamics of abusive relationships, (c) not identifying as a stereotypical IPV victim, (d) fear of reinforcing racial stereotypes, (e) internalizing social scripts regarding relationships, (f) structural barriers, and (g) leaving takes time. Twitter messages have the capacity to function as micronarratives that recount the complex barriers IPV victims confront when negotiating their relationships. This analysis provides a multifaceted description of the challenges associated with leaving abusive relationships that can augment existing theoretical frameworks on victim readiness. Furthermore, these findings demonstrate the myriad ways that societal representations of domestic violence (DV) serve as impediments for victims leaving their abusive relationships. Therefore, social media has the potential to provide a platform for capturing the lived experience of IPV.

Effects of PM2.5 on People's Emotion: A Case Study of Weibo (Chinese Twitter) in Beijing

PM_2.5 not only harms physical health but also has negative impacts on the public’s wellbeing and cognitive and behavioral patterns. However, traditional air quality assessments may fail to provide comprehensive, real-time monitoring of air quality because of the sparse distribution of air quality monitoring stations. Overcoming some key limitations of traditional surface monitoring data, Web-based social media platforms, such as Twitter, Weibo, and Facebook, provide a promising tool and novel perspective for environmental monitoring, prediction, and evaluation. This study aims to investigate the relationship between PM_2.5 levels and people’s emotional intensity by observing social media postings. This study defines the “emotional intensity” indicator, which is measured by the number of negative posts on Weibo, based on Weibo data related to haze from 2016 and 2017. This study estimates sentiment polarity using a recurrent neural networks model based on LSTM (Long Short-Term Memory) and verifies the correlation between high PM_2.5 levels and negative posts on Weibo using a Pearson correlation coefficient and multiple linear regression model. This study makes the following observations: (1) Taking the two-year data as an example, this study recorded the significant influence of PM_2.5 levels on netizens’ posting behavior. (2) Air quality, meteorological factors, the seasons, and other factors have a strong influence on netizens’ emotional intensity. (3) From a quantitative viewpoint, the level of PM_2.5 varies by 1 unit, and the number of negative Weibo posts fluctuates by 1.0168 units. Thus, it can be concluded that netizens’ emotional intensity is significantly positively affected by levels of PM_2.5. The high correlation between PM_2.5 levels and emotional intensity and the sensitivity of social media data shows that social media data can be used to provide a new perspective on the assessment of air quality.

Comparing the Use of Spatially Explicit Indicators and Conventional Indicators in the Evaluation of Healthy Cities: A Case Study in Shenzhen, China

Various indicator systems have been developed to monitor and assess healthy cities. However, few of them contain spatially explicit indicators. In this study, we assessed four health determinants in Shenzhen, China, using both indicators commonly included in healthy city indicator systems and spatially explicit indicators. The spatially explicit indicators were developed using detailed building information or social media data. Our results showed that the evaluation results of districts and sub-districts in Shenzhen based on spatially explicit indicators could be positively, negatively, or not associated with the evaluation results based on conventional indicators. The discrepancy may be caused by the different information contained in the two types of indicators. The spatially explicit indicators measure the quantity of the determinants and the spatial accessibility of these determinants, while the conventional indicators only measure the quantity. Our results also showed that social media data have great potential to represent the high-resolution population distribution required to estimate spatially explicit indicators. Based on our findings, we recommend that spatially explicit indicators should be included in healthy city indicator systems to allow for a more comprehensive assessment of healthy cities.

Social Profiling: A Review, Taxonomy, and Challenges

Social media has taken an important place in the routine life of people. Every single second, users from all over the world are sharing interests, emotions, and other useful information that leads to the generation of huge volumes of user-generated data. Profiling users by extracting attribute information from social media data has been gaining importance with the increasing user-generated content over social media platforms. Meeting the user's satisfaction level for information collection is becoming more challenging and difficult. This is because of too much noise generated, which affects the process of information collection due to explosively increasing online data. Social profiling is an emerging approach to overcome the challenges faced in meeting user's demands by introducing the concept of personalized search while keeping in consideration user profiles generated using social network data. This study reviews and classifies research inferring users social profile attributes from social media data as individual and group profiling. The existing techniques along with utilized data sources, the limitations, and challenges are highlighted. The prominent approaches adopted include Machine Learning, Ontology, and Fuzzy logic. Social media data from Twitter and Facebook have been used by most of the studies to infer the social attributes of users. The studies show that user social attributes, including age, gender, home location, wellness, emotion, opinion, relation, influence, and so on, still need to be explored. This review gives researchers insights of the current state of literature and challenges for inferring user profile attributes using social media data.

Crowdbreaks: Tracking Health Trends Using Public Social Media Data and Crowdsourcing

In the past decade, tracking health trends using social media data has shown great promise, due to a powerful combination of massive adoption of social media around the world, and increasingly potent hardware and software that enables us to work with these new big data streams. At the same time, many challenging problems have been identified. First, there is often a mismatch between how rapidly online data can change, and how rapidly algorithms are updated, which means that there is limited reusability for algorithms trained on past data as their performance decreases over time. Second, much of the work is focusing on specific issues during a specific past period in time, even though public health institutions would need flexible tools to assess multiple evolving situations in real time. Third, most tools providing such capabilities are proprietary systems with little algorithmic or data transparency, and thus little buy-in from the global public health and research community. Here, we introduce Crowdbreaks, an open platform which allows tracking of health trends by making use of continuous crowdsourced labeling of public social media content. The system is built in a way which automatizes the typical workflow from data collection, filtering, labeling and training of machine learning classifiers and therefore can greatly accelerate the research process in the public health domain. This work describes the technical aspects of the platform, thereby covering the functionalities at its current state and exploring its future use cases and extensions.

Ontology-Based Approach to Social Data Sentiment Analysis: Detection of Adolescent Depression Signals

BACKGROUND

Social networking services (SNSs) contain abundant information about the feelings, thoughts, interests, and patterns of behavior of adolescents that can be obtained by analyzing SNS postings. An ontology that expresses the shared concepts and their relationships in a specific field could be used as a semantic framework for social media data analytics.

OBJECTIVE

The aim of this study was to refine an adolescent depression ontology and terminology as a framework for analyzing social media data and to evaluate description logics between classes and the applicability of this ontology to sentiment analysis.

METHODS

The domain and scope of the ontology were defined using competency questions. The concepts constituting the ontology and terminology were collected from clinical practice guidelines, the literature, and social media postings on adolescent depression. Class concepts, their hierarchy, and the relationships among class concepts were defined. An internal structure of the ontology was designed using the entity-attribute-value (EAV) triplet data model, and superclasses of the ontology were aligned with the upper ontology. Description logics between classes were evaluated by mapping concepts extracted from the answers to frequently asked questions (FAQs) onto the ontology concepts derived from description logic queries. The applicability of the ontology was validated by examining the representability of 1358 sentiment phrases using the ontology EAV model and conducting sentiment analyses of social media data using ontology class concepts.

RESULTS

We developed an adolescent depression ontology that comprised 443 classes and 60 relationships among the classes; the terminology comprised 1682 synonyms of the 443 classes. In the description logics test, no error in relationships between classes was found, and about 89% (55/62) of the concepts cited in the answers to FAQs mapped onto the ontology class. Regarding applicability, the EAV triplet models of the ontology class represented about 91.4% of the sentiment phrases included in the sentiment dictionary. In the sentiment analyses, "academic stresses" and "suicide" contributed negatively to the sentiment of adolescent depression.

CONCLUSIONS

The ontology and terminology developed in this study provide a semantic foundation for analyzing social media data on adolescent depression. To be useful in social media data analysis, the ontology, especially the terminology, needs to be updated constantly to reflect rapidly changing terms used by adolescents in social media postings. In addition, more attributes and value sets reflecting depression-related sentiments should be added to the ontology.